Intelligent vehicles need to detect new classes of traffic objects while keeping the performance of old ones. Deep convolution neural network (DCNN) based detector has shown superior performance, however, DCNN is ill-equipped for incremental learning, i.e., a DCNN based vehicle detector trained on traffic sign dataset will catastrophic forget how to detect vehicles. In this paper, we propose a novel method to alleviate this problem, our key insight is that the original class of objects also appears in new task data, by utilizing these objects, our method effectively keeps the detection accuracy of original models while incremental learning to detect new classes of objects. Detailed experiments on PASCAL VOC dataset and TSD-max database verified the effectiveness of our method.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Learn to Detect Objects Incrementally


    Contributors:
    Guan, Linting (author) / Wu, Yan (author) / Zhao, Junqiao (author) / Ye, Chen (author)


    Publication date :

    2018-06-01


    Size :

    3187171 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    LEARN TO DETECT OBJECTS INCREMENTALLY

    Guan, Linting / Wu, Yan / Zhao, Junqiao et al. | British Library Conference Proceedings | 2018


    Train control, incrementally

    Foran, Pat | IuD Bahn | 2006


    INCREMENTALLY ADJUSTABLE SEAT ASSEMBLY

    ZOUZAL WINSEN C / PATRICK GERALD | European Patent Office | 2018

    Free access

    Incrementally Variable High-Voltage Supply

    Potter, D. W. / Chin, J. / Anderson, H. R. et al. | NTRS | 1985


    On causality observed incrementally, finally

    Mowbray, Miranda / Ferrari, Gian Luigi / Montanari, Ugo | TIBKAT | 1990